Our rapidly aging population will result in an increasing number of people at risk for loss of independence through dementia, frailty and other syndromes of aging. The high cost of health care to assist the dependent elderly is expected to soar by 2030, when over 20% of the U.S. population will be over the age of 65, many of whom will be unable to live independently. Some of this cost can be avoided by helping individuals remain independent and at home for as long as possible. This is an important goal by itself, as it is generally accepted that quality of life is higher for people who can remain in their own homes. The long-term objective of our research is to develop systems that improve our ability to unobtrusively monitor important health changes due to chronic disease and aging, allowing timely intervention to prevent avoidable health deterioration or loss of independence. It has been suggested that evolving sensor and other technologies provide a means of early detection and intervention minimizing morbidity and cost. However, the impact of such technologies is not well understood, and the hypothesis that data from these systems will allow older adults to remain independent is untested. The proposed project will evaluate how such technologies may contribute to care transition interventions. We will specifically examine real time patterns of mobility, sleep, medication use, weight change and social engagement to identify trends that may indicate compromised ability to maintain independence. Using technologies that we have developed and tested over the past five years in hundreds of seniors' homes, we will determine if such Ambient Independence Measures (AIMs), collected using unobtrusive sensors distributed throughout an individual's natural living environment, aid in making decisions about transitions to different levels of care. Our specific aims are: 1) Develop new algorithms (combining behavioral and physiological data) to identify trends in AIMs data that indicate acute or chronic changes that may compromise independence and conduct analyses to determine what AIMs data is of most value to care transition professionals; 2) Develop an automated system for presenting AIMs data to care teams on an as-needed basis, providing an early warning of the need for clinical follow-up. As part of this aim, we will create a shareable resource that facilitates the use of AIMs and related technology, allowing others to use this approach in their research or care-giving settings; and 3) Validate the AIMs metrics in senior community settings by determining if the use of the AIMs data assists caregivers in making decisions about transitions from independent living to higher care levels and changes the outcomes of those decisions.